Opioid addiction has remained widespread throughout the United States since the 1960s and a large proportion of users are involved in crimes to support their habits. After release from incarceration, relapse to opioid addiction is very common and this leads to more crimes and re- incarceration. Treatment advances in the area of medications have not reached this population. Effective medications such as methadone and buprenorphine are not well accepted by prosecutors and judges. Permission to conduct research on the most effective treatment approaches is very difficult to obtain for patients under legal restraint because informed consent is problematic. Naltrexone, an opiate receptor antagonist, has demonstrated pharmacological efficacy in preventing relapse to opioid addiction and it has been reported to be clinically effective in parolee populations although it is rarely used. Recently a depot formulation with a one month duration has received FDA approval for the treatment of alcoholism. The purpose of this study is to determine whether a monthly injection of naltrexone is practical and useful in the prevention of relapse and when compared to treatment as usual. We will also monitor HIV risk behaviors to determine whether the intervention reduces risk of HIV and hepatitis C infections. This collaborative project will take place in six treatment sites where there is a large population of parolees with a history of opiate addiction. In order to prevent even a subtle form of coercion, referrals from parole officers will not be accepted. After determining that all volunteers are opiate free by urine test and not currently opiate dependent using a naloxone test, they will be randomized to depot naltrexone or Treatment as Usual (TAU). Participants in both groups will be given identical follow up monthly for six months with measures of opiate use by self-report, urine test and hair analysis. An additional random urine test will take place each month between monthly visits. Both groups will be re-evaluated six and 12 months later. The University of Pennsylvania will be the coordinating site and each site will have a randomization goal of 20 new patients per year over 3.5 to 4 years to accrue a total of 360 to 400 participants. Treatment outcome will be measured by urine tests, hair analysis, self-report and continuation in treatment. Both naltrexone and comparison groups will receive equivalent voucher incentives to remain in the program. A benefit-cost analysis will be conducted to compare the costs of the treatment with the quantifiable benefits in terms of reduced crime, re-incarceration and medical services and increased employment.